Proc. Nail. Acad. Sci. USAVol. 75, No. 2, pp. 735-739, February 1978Biochemistry

Human mitochondrial DNA: Analysis of 7S DNA fromthe origin of replication

(gel electrophoresis/restriction endonucleases/electron microscopy/hybridization/multiple components)

WESLEY M. BROWN, JOHN SHINE, AND HOWARD M. GOODMANDepartment of Biochemistry and Biophysics, University of California, San Francisco, California 94143

Communicated by Norman Davidson, November 17, 1977

ABSTRACT Heat-treated samples of human mitochondrialDNA (mtDNA) exhibited a set of three low molecular weightDNA bands in addition to the major mtDNA band when elec-trophoresed in polyacrylamide gels. These DNA componentswere seen only after heat treatment or after relaxation of themtDNA with a restriction endonuclease. The three componentswere single stranded and had sizes of 550, 585, and 620 nu-cleotides, close to the size (600 nucleotides) estimated fromcontour length measurements for the 7S DNA from the D loopof human mtDNA. Hybridization of the components with re-striction endonuclease fragments of known position in themtDNA confirmed this identification. Digestion of each 7SDNA component with the restriction endonuclease Hae IIIproduced three fragments, two of which were identical in sizeamong the components and the third of which varied. This thirdfragment, shown to be from the 5' end of each component, dif-fered in length by -35 nucleotides among the components.These results suggest that human 7S mtDNA synthesis is ter-minated at a distinct position and that it is either initiated atone of three possible sites in the same mtDNA or that themtDNA population consists of three subpopulations, each dif-fering from the others by the presence or absence of a nucleotidesequence immediately adjacent to the origin of replication.

The human mitochondrial genome is a closed circular duplexDNA molecule of 16,500 E 300 base pairs that is replicatedasymmetrically and unidirectionally from a fixed point. Someanimal mitochondrial DNAs (mtDNAs), including humanmtDNA, possess a characteristic structure at the origin of rep-lication, the D loop, that can be seen by electron microscopy(1). D loop formation occurs via the displacement of one of thestrands (the H strand) of mtDNA in the region of the origin ofreplication by the synthesis of a short complementary piece ofDNA, referred to as 7S DNA, the synthesis of which is termi-nated by an unknown mechanism. As shown by 7S DNAturnover (2,3), extension of the 7S DNA and synthesis of DNAcomplementary to the displaced strand both occur more slowlythan D loop formation. This gives rise to a subpopulation of themtDNA that consists of molecules having D loops of a uniformsize. These molecules co-band with closed circular DNA thatdoes not contain a D loop in a CsCl/ethidium bromide buoyantdensity gradient. Molecules in later stages of replication bandat positions intermediate between the closed circular (lowerband) and open circular plus linear (upper band) DNA (4). Thisallows the simultaneous purification and separation of D loopmtDNA from other replicating forms.

In this report we describe the dissociation of the 7S DNAfrom human mtDNA preparations and the analysis of this DNAby restriction endonuclease digestion and gel electrophoresis.

The implications of the results on our concept of the structureof the origin of replication of human mtDNA and the mecha-nism of D loop formation are discussed.

MATERIALS AND METHODSDNAs. mtDNAs were prepared as described (5, 6). The

sources of mtDNA were: human, Heba (S3) cells and placentas;green monkey (Cercopithecus aethiops), BSC-1 cells; talapoinmonkey (Miopithecus talapoin), primary cells; woolly monkey(Lagothrix cana), primary cells; and mouse (Mus musculus),LA9 cells. The cell culture conditions used have been described(5, 6). Bacteriophage PM2 DNA, prepared by the method ofEspejo et al. (7), was a gift from Bob Watson. BacteriophageM13 DNA digested with the restriction endonuclease Hae IIIwas a gift from Peter Seeburg.

Electron Microscopy of DNA. DNA in 50% formamide wasspread by the procedure of Davis et al. (8). Estimates of the Dloop frequency were obtained by scoring 200-300 moleculesper sample. Contour length measurements were performed asdescribed (6).Gel Electrophoresis of DNA. The electrophoresis of DNA

in polyacrylamide slab gels (9) and in agarose slab gels (10) hasbeen described. Gels were either photographed after stainingwith ethidium bromide or, for 32P-labeled samples, dried onWhatman 3MM paper and autoradiographed with KodakNS-2T x-ray film. The sizes of DNA fragments in the gels wereestimated by use of PM2 duplex DNA digested with restrictionendonuclease HindIII (11, 12) and M13 single-stranded DNAdigested with restriction endonuclease Hae III (13) as sizestandards.

Preparation of 7S DNA. Kasamatsu et al. (1) have describedthe small, noncovalently bound DNA found in the D loop ofmouse mtDNA as 7S DNA, according to its sedimentationcoefficient. We have designated the corresponding DNA frommtDNAs of different species as 7S DNA, but no correspondencewith the actual sedimentation coefficient is implied. Unlessotherwise noted, 7S DNA was dissociated from D loop mtDNAby heating at 80° in 50 mM NaCl/10mM Tris, pH 7.8/1 mMEDTA for 3-5 min. The 7S DNA was separated from themtDNA by electrophoresis through a polyacrylamide (0.2% bis)slab gel at 100 V for 3-4 hr (room temperature). After elec-trophoresis of human mtDNA, three 7S bands were seen, des-ignated A, B, and C in order of decreasing size. For some ex-periments the individual bands were cut out of the gel and theDNA was extracted as described (14).Enzymes and Enzymatic Reactions. Ribonuclease-free

bacteriophage T4 polynucleotide kinase, prepared by the

Abbreviations: mtDNA, mitochondrial DNA; Hepes, 4-(2-hydroxy-ethyl)-1-piperazineethanesulfonate.

735

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby marked"advertisement" in accordance with 18 U. S. C. §1734 solely to indicatethis fact.

Proc. Natil. Acad. Sci. USA 75 (1978)

method of Panet et al. (15), was a gift of Herb Heyneker. 1jy-32PjATP was prepared and used with polynucleotide kinase tolabel the 5' termini of DNA molecules as described (14). Ter-minal transferase was a gift of R. L. Ratliff. The reaction con-ditions are given in the following section. The restriction en-donucleases HincII, Kpn I, Hae III, and Hha I were purchasedfrom New England Biolabs (Beverly, MA). These enzymes andtheir respective reaction conditions are described in ref. 16.

Preparation of Labeled 7S DNA with High Specific Ac-tivity. A 7S DNA of high specific activity was prepared bysynthesizing poly([32P]dT) tails on the 3' termini of the 7S DNAwith terminal transferase. A 30-Al mixture containing -1 ngeach of the 7S DNA components A, B, and C in 0.2 M 4-(2-hydroxyethyl)-1-piperazineethanesulfonate (Hepes), pH7.5/400 ,g of bovine serum albumin per ml/0.5 mM 2-mer-captoethanol/1 mM CoCl2 was chilled to 00, then added to atube containing 0.1 mCi of dried [a-32P]dTTP (Amersham).Terminal transferase (1 ul, 250 units) was added and the mix-ture placed at 37°. The reaction was monitored at 5-min in-tervals and stopped after 20 min by the addition of 1 gl of 0.33MEDTA (pH 8.5) and 1 Ml (10 Mg) of carrier tRNA. The total in-corporation of 32P into acid-insoluble material was 3 X 106 cpm,or -1 X 109 cpm/Mg of DNA. This level of activity is routinelyobtained and the contribution of spontaneously initiated syn-thesis of poly(dT) is small (-100%) if the reaction is stopped after-20 min (H. J. Kung, personal communication).An equal volume of 0.5 M NaCI/0.2% sodium dodecyl sul-

fate/10 mM Tris (pH 7.8)/1 mM EDTA was added to the re-action mixture. The labeled DNA probe was separated fromunincorporated [a-32PJdTTP on a 0.6 X 15-cm Sephadex G50column, using the same buffer for the elution. The fractionscontaining the DNA probe were pooled and extracted once withphenol/chloroform. The aqueous phase was made 0.3 M insodium acetate (pH 5.8) and precipitated with ethanol. Aftercentrifugation and removal of the ethanol the DNA probe wasresuspended in 100 ML of 10 mM Tris (pH 7.8)/1 mM EDTA.Based on a total incorporation of 3 X 106 cpm, the recovery atthis stage was 30% (9 X 105 cpm).

Hybridization of 7S DNA to mtDNA Restriction Frag-ments. Restriction endonuclease digests of HeLa mtDNA withKpn I, HinclI, and Hha I were electrophoresed in a 2% agaroseslab gel for 2.5 hr at 100 V. HindIII fragments of PM2 DNA,labeled at their 5' ends with 32p, were included as size standards.The gel was stained and photographed; then the DNA in thegel was transferred to a nitrocellulose filter and immobilized(17).The high specific activity 7S DNA was diluted to a volume

of 4 ml with hybridization buffer [50 mM Hepes (pH 7.5)/3mM EDTA/500,Mg of yeast carrier RNA per ml/]/5% sodiumdodecyl sulfate/0.45 M NaCl/45 mM Na citrate (pH 7.5)/50%formamidel. The nitrocellulose filter was placed on SaranWrap, saturated with the hybridization mixture, sealed in SaranWrap and aluminum foil, and placed at 370 for 60 hr. The filterwas washed for 60 min in 500 ml of 0.3 M NaCI/30 mM Nacitrate at pH 7.5, at 250, and for 18 hr in 500 ml of 0.9 MNaCl/90 mM Na citrate at pH 7.5, with four buffer changes,after which the filter was air dried and then autoradiographedwith Kodak NS-2T x-ray film.

5'-End Analysis. DNA labeled at the 5' ends with 32p wasdigested to 5'-mononucleotides by incubation at 370 for 3 hrwith 200 Mug each of DNase I and venom phosphodiesterase(Worthington) per ml in 100 mM Tris (pH 8.5)/50mM MgCI2.The 5'-dNMPs were separated by high-voltage paper electro-phoresis in 5% acetic acid/0.5% pyridine/10mM EDTA at pH3.5.

a0 ~~0

.4I*. + rt'-

FIG. 1. Polyacrylamide gel electrophoresis of 7S DNA dissociatedfrom human placenta mtDNA by introducing double-stranded scis-sions (lane 1) and by heating (lane 2). The 7S DNA migrates as threecomponents, designated A, B, and C in order of decreasing size. Thedouble-stranded scissions were produced by the restriction endonu-clease Kpn I, which does not cleave human mtDNA within the D loop(unpublished data).

RESULTSFrequency and Size of D Loops in Human mtDNA. Elec-

tron microscopy was used to estimate the frequency of D loopsin molecules isolated from lower band preparations of humanmtDNAs. The D loop frequency in five different preparationsof HeLa cell mtDNA averaged 8% (range: 5-12%), in agree-ment with the value, 7%, published by Kasamatsu et al. (18).Much higher frequencies (53 and 57%) were observed inmtDNA preparations from two human placentas.The average contour length of the duplex branch of the D

loop, expressed as a fraction of the total contour length of themtDNA molecule, was 0.036 ± 0.004 SD (54 HeLa mtDNAmolecules) or 595 + 65 nucleotides, based on an estimated ge-nome size of 16,500 base pairs. For 12 placenta mtDNA mol-ecules the value was 0.037 ± 0.006, a value not significantlydifferent from the value for the HeLa D loops.

Dissociation of 7S DNA from mtDNA. 7S DNA may bedissociated from mtDNA by heating (18) and by the intro-duction of single- or double-strand scissions (2, 6). Humanplacenta mitochondrial 7S DNA was prepared by each of thesemethods and compared by polyacrylamide gel electrophoresisafter 5'-end labeling with 32P. In both cases the 7S DNA mi-grated as three distinct bands of equal intensity, designated A,B, and C (Fig. 1). The darkening of the negative in the lowerthird of lane 2 is presumably due to mitochondrial RNA speciescomplexed with the mtDNA until their release by the heattreatment. Their absence in lane 1 (except at the extreme bot-tom) is probably due to a small amount of RNase activitypresent in the Kpn I enzyme preparation.

Sizes of Human 7S DNA Components. The A, B, and Ccomponents of human placenta mitochondrial 7S DNA wereseparated by polyacrylamide gel electrophoresis. The regionof the gel containing each component was excised and the DNAwas eluted from it. A portion of each component was re-lec-trophoresed along with an aliquot from the original mixture.

736 Biochemistry: Brown et al.

Proc. Natl. Acad. Sci. USA 75 (1978) 737

1 2 3 4 5

FIG. 2. Polyacrylamide gel electrophoresis of the individual A,B, and C components of 7S DNA isolated from human placentamtDNA by preparative polyacrylamide gel electrophoresis. Lanes 1-5contain HindIII-digested PM2 DNA, 7S-A, 7S-B, 7S-C, and the un-separated 7S DNA components, respectively.

Each component ran according to its original mobility (Fig. 2)and each was -95% free of contamination by the other twocomponents. The mobilities of the components were also un-

changed by annealing at 24° below the Tm for 14 hr and byincubation at high pH (data not shown). These results are

consistent with the interpretation that the differences in theelectrophoretic mobilities of the components are due only tosize differences.The sizes of the components were estimated by polyacryl-

amide gel electrophoresis, with the Hae III fragments of thesingle-stranded DNA from bacteriophage M13 as size standards(13). The sizes obtained for the A, B, and C components were

680, 645, and 615 nucleotides, respectively. These size estimatesare within experimental error but slightly larger than those forthe size of the D loop of human mtDNA estimated by contourlength measurements. We regard the sizes from contour lengthmeasurements as the more reliable, since the sizes of the HaemI, M13 fragments are not known with precision. However, thesize differences among the components, 30-5 bases, are lesssensitive to error than the absolute sizes, and we consider theseestimates to be reliable.

Multiple 7S Components in Other mtDNAs. The occur-rence of multiple 7S components in mtDNAs is not an isolatedphenomenon, but appears to be ubiquitous, at least amongmammalian species. The electrophoretic mobilities of 7S DNAcomponents from three different human placenta mtDNApreparations, two HeLa mtDNA preparations, and mtDNApreparations from four other mammalian species were com-

pared. Of the two HeLa mtDNA preparations used, one camefrom the lower band of a CsCl/ethidium bromide buoyantdensity gradient and the other from the portion of the gradientbetween the lower and upper band positions.The human samples (Fig. 3, lanes 1-4) all exhibited identi-

cally migrating A, B, and C bands. The two HeLa samples (Fig.3, lanes 3 and 4) showed differences in the relative intensitiesof the bands, the band corresponding to the C component beingless intense than the other two bands. The HeLa sample from

2 3 4 5 6 7 .8

FIG. 3. Polyacrylamide gel electrophoresis of heated mtDNAs

from two human placenta samples (lanes 1 and 2), HeLa cells (lanes

3, lower band, and 4, intermediate; see text), green monkey (lane 5),

talapoin monkey (lane 6), woolly monkey (lane 7), and mouse (LA9)

cells (lane 8). The slanted bands in lanes 1 and 2 are due to uneven

drying of the gel prior to autoradiography. The heavy band (labeled

mt) that appears in all lanes is due to a fast migrating portion of

full-length linear duplex mtDNA (see text).

the intermediate region of the gradient (Fig. 3, lane 4) showedan additional fast component not seen in the other samples. Theband indicated by the arrow in Fig. 3, common to all of thesamples, has been shown to correspond to full-length linearduplex mtDNA, most of which does not enter the gel (W.Brown and F. DeNoto, unpublished results).mtDNA samples from green monkey (lane 5), talapoin

monkey (lane 6), woolly monkey (lane 7), and mouse (LA9) cells(lane 8) all exhibited multiple 7S components (Fig. 3). Thesediffered in both number and position from the human 7Scomponents, but their estimated sizes fell within a narrow range(450-700 bases).

Hybridization of 7S DNA to Restriction EndonucleaseFragments of mtDNA. To establish that the 7S DNA cameexclusively from the D loop of the mtDNA, we hybridized 7SDNA to restriction endonuclease fragments of mtDNA. Digestsof HeLa mtDNA with the restriction endonucleases Kpn I,HinclI, and Hha I were electrophoresed in a 2% agarose gel(Fig. 4a). The DNA was denatured and transferred onto a ni-trocellulose filter as described (17). A high-specific-activity 7SDNA was prepared and hybridized with the filter-bound DNAas described in Materials and Methods. After hybridization,the filter was autoradiographed. The result (Fig. 4b) shows thatthe 7S DNA hybridized to the smaller Kpn I fragment and tothe largest fragments in each of the HincIl and Hha I digests.From restriction endonuclease cleavage mapping studies (un-published data), we know that the D loop of human mtDNAoccurs on these same three fragments and, furthermore, thatthe only portion of the mtDNA common to the three fragmentsis from 0.956 to 0.040 genome units. Since the origin of repli-cation is at 0 and the D loop extends to 0.036 genome units, thisis good evidence that the 7S DNA corresponds to the single-stranded DNA segment present in the D loop. The 7S DNA usedin the hybridization contained all three components; thus thisresult also indicates that the components originate from thesame portion of the genome. Sequence data (unpublished re-sults) confirm this conclusion.

Examination of Fig. 4b also indicates that a significantamount of the 7S DNA hybridized with the band containingthe larger Kpn -I fragment. This could have been due, in part,to the presence- in this band of incompletely digested linearmtDNA molecules. However, it is more likely that the observedhybridization was with the smaller Kpn I fragments exclusively,

Biochemistry: Brown et al.

738 Biochemistry: Brown et al.

A..~~~

i:-Was:

l 2 3 4-A _ . _W.b.._ _ ~~~~~~~~~~~,qE_: 40

a

FIG. 4. (a) Agarose gel electrophoresis ofKpn I, HincII, and HhaI digests of HeLa mtDNA. (b) Hybridization of a 32P-labeled 7S DNAprobe to the DNA fragments shown in a. The four lanes shown in botha and b are (from left to right) HindIII-digested PM2 DNA and KpnI, HincIl, and Hha I digests of HeLa mtDNA. The HindIll-digestedPM2 DNA fragments were labeled at the 5' ends before electropho-resis.

and that some of these were "trapped" in the larger fragmentband of the gel. The "trapping" of smaller DNA fragments inbands containing larger DNA fragments is, in our experience,a routine feature of DNA electrophoresis in both agarose andpolyacrylamide gels (see also Figs. 5 and 6).

Differences among 7S DNA Components. The mitochon-drial 7S DNA components were labeled at their 5' ends anddigested with the restriction endonucleases Hae III and HhaI, both of which cleave single-stranded DNA (19, 20). The re-

sults (Fig. 5) indicate that Hha I does not cleave any component,that Hae III cleaves all three, and that at least some of the

'a-s.

-

1 2 3 4 5 6 7 8 9 10

FIG. 5. Polyacrylamide gel electrophoresis of 5'-end-labeledcomponents of human placenta 7S mtDNA digested with Hae III andHha I restriction endonucleases. The lanes contain: 1, HindIII-di-gested PM2 DNA; 2 and 10, Hae III-digested M13 DNA; 3,4, and 5,Hae III-digested 7S components A, B, and C, respectively; 6, undi-gested (control) heated mtDNA; 7, 8, and 9, Hha I-digested 7S com-ponents A, B, and C, respectively. In each of lanes 3,4, and 5 there isone faint band (arrows) that migrates more slowly than the main band.This faint band represents an incompletely digested Hae III fragmentwith a size -50 nucleotides greater than the main band, in each case,and indicates that there is a Hae III fragment of -50 nucleotidesimmediately adjacent to the 5'-terminal Hae III fragment.

-450

FIG. 6. Polyacrylamide gel electrophoresis of components ofhuman placenta 7S mtDNA digested with Hae III restriction endo-nuclease and then 5'-end labeled. The lanes contain: 1 and 5, Hin-dIII-digested PM2 DNA; 2,3, and 4, Hae III-digested 7SDNA com-ponents A, B, and C, respectively. Single-stranded fragment sizes (innucleotides) are shown at the right. The faint bands indicated by thecurved arrows and the faint bands at a size of 370 nucleotides are in-complete Hae III digestion products.

variation in the electrophoretic mobilities of the componentsis due to a difference in size at or near the 5' ends. The estimatedsizes of the 5'-terminal Hae III fragments bf the A, B, and Ccomponents are 232, 198, and 168 bases, respectively.To determine if the size differences resided exclusively at or

near the 5' ends, we digested the three components with HaeIII and then labeled them at their 5' ends. Each component wascleaved into three pieces by Hae III (Fig. 6). The largest (-320bases) and smallest (-50 bases) fragments did not vary in sizeamong the three digests. The incompletely digested fragments,present in low amounts and noted in Figs. S and 6, revealed thatthe smallest Hae III fragment is the internal piece, since it islinked to both the intermediate-sized fragment, shown to bethe 5' end, and the largest fragment, which must correspondto the 3' end. Thus, the size variation among the 7S DNAcomponents resides entirely within the first 170-230 bases ofthe 5' termini (Fig. 7).With three exceptions, the faint bands seen in Figs. 5 and 6

5' Hae IIIIIT

I

3'

-A

-B

c

FIG. 7. Locations of the Hae III cleavage sites in the 7S DNAcomponents. The horizontal scale is graduated in units of 50 nucleo-tides. The vertical arrows show the positions of the Hae III cleavagesites.

5

-4370-'4320

4232

-4198

-4168

Proc. Nati. Acad. Sci. USA 75 (1978)

Proc. Natl. Acad. Sc. USA 75 (1978) 739

Table 1. 5'-End analysis of mitochondrial 7S DNA components

% of total 32p, cpm Total 32p,DNA Origin dCMP dAMP dGMP TMP cpm

7S-A 2.3 6.9 17.1 24.5 49.2 32267S-B 2.3 4.8 46.4 32.3 14.3 24957S-C 10.6 25.7 6.0 25.4 32.4 2083HindIII, PM2 0.4 0.4 99.3 0.0 0.0 2297

may be explained by incomplete digestion and by "trapping."One exception is the band in lane 2 of Fig. 6, at a positionequivalent to -215 nucleotides, which also occurs in lane 3 ofFig. 5. The other exceptions are the two bands seen in lane 4 ofFig. 6, between the main bands at 168 and 50 nucleotides. Thesefaint bands are noted but not understood.

5'-Terminal Nucleotides of 7S DNA Components Are NotUnique. The 7S DNA components were labeled at their 5' endswith 32P and digested to 5'-mononucleotides with pancreaticDNase and venom phosphodiesterase, and the mononucleotideswere separated by high-voltage electrophoresis. The results ofone set of determinations are given in Table 1. Although the5'-end analysis of the HindIII-digested PM2 DNA control gavethe expected result (HindIII cleavage produces exclusively5'-dA fragments), the 7S components showed no unique 5' endfor any component. Determinations on two other 7S DNAsamples obtained from different placentas showed similar 5'heterogeneity, but yielded percent base compositions differentfrom each other and from those in Table 1. Whether this latterdifference is biologically significant or, e.g., an artifact ofpreparation, is unknown. Heterogeneity at the 5' end of the 7SDNA from mouse (LA9) mtDNA has also been observed (18).Because of this it has not been possible to obtain the base se-

quences at the 5' ends of the human 7S DNA components.

DISCUSSIONThe results indicate that the 7S mtDNA corresponds to thesmall, noncovalently bound DNA fragment found in the D loopof human mtDNA by the criteria of size and specificity of hy-bridization. The lack of bands that migrate more slowly thanthe A component of the 7S DNA in the HeLa sample from theintermediate region of the CsCl/dye gradient (lane 4, Fig. 3)suggests that the stepwise synthesis of mtDNA is confined tothe formation of the 7S species. The origin of the fast migratingband observed in this same sample has not been investi-gated.On the basis of the data from electron microscopy and hy-

bridization to restriction endonuclease fragments of mtDNAwe have assumed that the three human 7S DNA componentsrepresent the same mtDNA sequence, except at the 5' endswhere the components differ in size.We have recently determined the sequence of the first 33

bases from the 5' end of the smallest Hae III fragment of the7S DNA (unpublished data). This was done by pooling the A,B, and C components before Hae III digestion; hence the se-

quence is derived from the smallest Hae III fragment commonto each of the components. A sequence could not have beenobtained if this Hae III fragment differed among the threecomponents. From these data we conclude that the three 7SDNA components are probably identical in sequence, exceptfor a region at or near the 5' ends, and also that the 7S DNA iscomplementary to one and not to both of the human mtDNAstrands, a fact already known for mouse 7S mtDNA (1).The exact nature of the size variation at the 5' end of the

human 7S mtDNA is unknown. The most intriguing possibilityis that the size variation is produced by 7S DNA synthesis thatis initiated at any one of three sites in the mtDNA, located 30-35nucleotides apart. It is also possible that the three 7S DNAcomponents are produced on three classes of mtDNA molecules,each class differing from the other two by an added or deletedsequence of 30-35 nucleotides near the origin of replication.A third possibility is that the initiation point of human mtDNAsynthesis is not specified by a sequence, but occurs instead asa consequence of some other regional feature of the mtDNAor in relation to another macromolecular species that is associ-ated with the mtDNA. In this case the size variation among the7S components could be related to normal variation in su-perhelix density, and heterogeneity in the 5'-terminal nucleo-tide would be expected. This latter heterogeneity could also berelated to heterogeneity at the 3' end of a hypothetical prim-er.The lack of size variation at the 3' end of the human 7S DNA

does not support the hypothesis (18) that the termination of 7SDNA synthesis is due to superhelix contraints. It is possible thata specific sequence terminates the synthesis of 7S DNA in thehuman mitochondrial genome.

We thank Francis DeNoto for technical assistance, Harold Varmusfor helpful comments on the manuscript, and Tomi Gottschall for as-sisting in the preparation of the manuscript. This work was supportedby grants from the National Science Foundation (BMS75-10468) andNational Institutes of Health (CA 14026) and by a U.S. Public HealthService postdoctoral traineeship to W.M.B.

1. Kasamatsu, H., Robberson, D. L. & Vinograd, J. (1971) Proc. Nati.Acad. Sci. USA 68,2252-2257.

2. Robberson, D. L. & Clayton, D. A. (1973) J. Biol. Chem. 248,4512-4514.

3. Berk, A. J. & Clayton, D. A. (1974) J. Mol. Biol. 86,801-824.4. Kasamatsu, H. & Vinograd, J. (1974) Annu. Rev. Biochem. 43,

695-720.5. Brown, W. M. (1976) Ph.D. Dissertation, California Institute of

Technology.6. Brown, W. M. & Vinograd, J. (1974) Proc. Natl. Acad. Sci. USA

71,4617-4621.7. Espejo, R. T., Cannello, E. S. & Sinsheimer, R. L. (1968) Proc.

Natl. Acad. Sci. USA 63, 1164-1168.8. Davis, R. W., Simon, M. & Davidson, N. (1971) in Methods in

Enzymology, eds. Grossman, L. & Moldave, K. (Academic Press,New York), Vol. 21, pp. 413-428.

9. Maniatis, T., Jeffrey, A. & van deSande, V. H. (1975) Biochem-istry 14, 3787-3794.

10. Brown, W. M., Watson, R. M., Vinograd, J., Tait, K. M., Boyer,J. W. & Goodman, H. M. (1976) Cell 7,517-530.

11. Brack, C., Eberle, H/= Bickle, T. A. & Yuan, R. (1976) J. Mol.Biol. 104, 305-309.

12. Parker, R. C., Watson, R. M. & Vinograd, J. (1977) Proc. Natl.Acad. Sci. USA 74, 851-855.

13. Seeburg, P. H., Shine, J., Martial, J., Ullrich, A., Baxter, J. D. &Goodman, H. M. (1977) Cell 12, 157-165.

14. Maxam, A. M. & Gilbert, W. (1977) Proc. Natl. Acad. Sci. USA74,560-564.

15. Panet, A., van deSande, J. H., Loewen, P. C., Khorana, H. G.,Raae, A. J., Lillehaug, J. R. & Kleppe, K. (1973) Biochemistry 12,5045-5050.

16. Roberts, R. J. (1976) CRC Crit. Rev. Biochem. 3, 123-164.17. Southern, E. M. (1975) J. Mol. Biol. 98, 503-517.18. Kasamatsu, H., Grossman, L. I., Robberson, D. L., Watson, R. &

Vinograd, J. (1973) Cold Spring Harbor Symp. Quant. Biol. 38,281-288.

19. Horiuchi, K. & Zinder, N. D. (1975) Proc. Natl. Acad. Sci. USA72,2555-2559.

20. Blakesley, R. W. & Wells, R. D. (1975) Nature 257,421-422.

Biochemistry: Brown et al.